Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Changing Patterns of Psychiatric Inpatient Care in the United States, 1988-1994

Changing Patterns of Psychiatric Inpatient Care in the United States, 1988-1994 Using data from the National Hospital Discharge Survey and the Inventory of Mental Health Organizations, this article examines national trends in psychiatric inpatient care from 1988 to 1994 in general hospitals and mental hospitals. We find that discharges with a primary diagnosis of mental illness in general hospitals increased from 1.4 to 1.9 million during this period. The total increase of 1.2 million days of care in general hospitals was small relative to the reduction of 12.5 million inpatient days in mental hospitals. General hospital discharges increased most in private nonprofit hospitals and declined substantially in public hospitals. Length of stay has fallen most substantially in private nonprofit hospitals. Public programs have increasingly replaced private insurance as the major source of payment. These observations suggest that psychiatric inpatient care in general hospitals can be characterized as a process in which patients who would have been clients of public mental hospitals in a prior period replace privately insured patients who, under managed care, are largely treated in community settings. Private nonprofit general hospitals increasingly treat publicly financed patients with more severe illnesses.In recent years the continuing reduction of resident populations in long-term mental hospitals, hospital closures and mergers, managed care, and an increasingly competitive marketplace have transformed the psychiatric inpatient sector. Resident populations in public psychiatric hospitals fell to less than 80000 in the 1990s.In contrast, the number of specialized psychiatric units in general hospitals increased from 664 to 1516 between 1970 and 1992 and the number of private mental hospitals more than tripled, with inpatient admissions quadrupling.Most Americans are now in behavioral health care programsand most persons with severe mental illness (SMI) reside in the community, commonly with Medicaid coverage. Medicaid managed care enrollment has been growing rapidly, increasing from less than 10% in 1991 to 48% in 1997.The consequences of these changes are poorly understood.In this article we examine how these changes in the health care system have shaped patterns of psychiatric inpatient care. We explore the hypothesis that managed care reduces inpatient care for less seriously ill persons in general hospitals and that these patients are replaced by persons with more severe disorders who in earlier periods received their care in public mental hospitals. By examining trends in relationship to ownership we also seek insights into how competition affects selection into different types of hospitals. We do this by examining changes in discharge numbers, rates and days of care, lengths of stay, proportions of inpatients with SMI, sources of payment, and discharge status.MATERIALS AND METHODSData for these analyses come from the National Hospital Discharge Survey (NHDS) and the Inventory of Mental Health Organizations and General Hospital Mental Health Services (hereafter referred to as "Inventory") for the period 1988 to 1994.The NHDS, described in detail elsewhere,is a multistage sample of discharges from nonfederal hospitals with an average length of stay of less than 30 days. Short-stay mental hospitals are eligible for inclusion in the sample, but the number included is small and represents approximately 13% of all psychiatric discharges (R. Pokras, written communication, August 1997). For all analyses, sample statistics are weighted to allow for national estimates.Analyses are limited to discharges with first-listed psychiatric diagnoses coded between 290 and 319 according to the International Classification of Diseases, 9th Revision, Clinical Modification.Severe mental illness includes a first-listed diagnosis of schizophrenia and related disorders (codes 295, 297-299), bipolar disorder (296.0, 296.1, 296.4-296.9), obsessive-compulsive disorder (300.3), or major depressive disorder with psychotic features (296.34, 296.24). Rates of discharge are calculated using US Census estimates of the adult, civilian, noninstitutionalized population on July 1 of each year.We distinguish between public hospitals, including those controlled by state and local governments; private nonprofit hospitals, including those operated by not-for-profit organizations and churches; and proprietary hospitals, including those operated for profit by individuals, corporations, or partnerships. Changes in hospital ownership are coded every 3 years; thus, the 1990 data do not accurately reflect ownership for those hospitals that changed control after 1988. The number of hospitals miscoded in 1990, however, is likely to be small because only approximately 1% of hospitals change ownership each year.Other variables include a continuous measure of length of stay (difference between admission and discharge date) and a categorical variable measuring short lengths of stay, defined here as 5 days or less. Total days of care is the sum of lengths of stay. Expected payer source is defined as the primary listed source of payment. Transfer rates are the ratio of the number of patients transferred to another institution at discharge to the total number of discharges for a particular group.The Inventory, a biennial national survey by the Center for Mental Health Services (Rockville, Md) of facilities that provide specialty mental health care, provides supplementary data and is described in detail elsewhere.The Inventory included very limited clinical information for each admission, and we use it only to estimate changes in the number and rates of discharges and total days of care from mental hospitals. Sampling overlap between the NHDS and the Inventory may in some cases increase aggregate estimates by 10% to 15% when data from these 2 samples are combined.Because the Inventory is based on surveys of the universe of hospitals, tests for statistical significance are not appropriate. We use the parameters provided by the National Center for Health Statistics to calculate approximate SEs for various estimates for the NHDS.We highlight when estimates are based on less than 60 cases or have a relative SE of more than 30%, because these estimates may be unreliable. Calculation of the SE of percentages assumes that the relative SE of the denominator is less than 5% of the estimate, or that the relative SEs of both the denominator and the numerator are less than 10%. Some subgroup analyses presented here violate this assumption. However, the SEs and tests of significance presented probably will be conservative, given that we cannot calculate exact SEs for the major variables. We recommend that attention be focused on the magnitude of substantive differences between groups, and not solely on whether the test statistic attains statistical significance.Statistical inferences about differences between types of hospitals are made through 2-tailed ttests. Analyses of changes over time used weighted least squares regression,testing whether there is a linear relationship between time (year of survey) and each of our measures.RESULTSNUMBER AND RATE OF DISCHARGEAs presented in Table 1, between 1988 and 1994 the number of discharges for a primary psychiatric diagnosis from general hospitals significantly increased by approximately 35%, from 1.4 to 1.9 million discharges. The rate of discharges also significantly increased during this period, from 785 to 996 discharges per 100000 adult population. With a modest decrease in discharges during the same period for other illnesses (1.4%), psychiatric discharges increased as a proportion of all discharges from 5.1% to 6.8%.Table 1. Psychiatric Discharges (in Millions) From General Hospitals and Rate per 100000 Adult Population by Hospital Ownership*See table graphicThe increase in the number and rate of discharges has been confined to the private sector. The largest growth in inpatient psychiatric care in general hospitals occurred in private nonprofit hospitals (53% increase), with a smaller (29%), though not statistically significant, increase in the number of discharges from proprietary hospitals. While significantly more discharges were from public hospitals than proprietary hospitals in 1988, in 1994 the opposite pattern emerges. Indeed, during these 6 years public hospitals experienced an almost one-third reduction in number of discharges. Rates of discharge per 100000 population follow this same pattern.During this same period, discharges from mental hospitals also increased, although the rate of discharge from these hospitals, unlike that from general hospitals, remained relatively stable (Table 2). Discharges from public mental hospitals declined by approximately 20%, while increasing by more than 27% in private nonprofit and proprietary hospitals. Combining discharges from general and mental hospitals shows an aggregate increase of more than half a million discharges in 1994 as compared with 1988. During this period, the proportion of all discharges accounted for by general hospitals increased from 68% to 73%.Table 2. Discharges From Mental Hospitals and Rate per 100000 Adult Population by Hospital Ownership*See table graphicLENGTH OF STAYIn the period 1988 to 1994, psychiatric length of stay in general hospitals declined significantly from 12.1 to 9.6 days. As shown in Figure 1, the steepest decline occurred in nonprofit general hospitals, where average length of stay fell from 12.6 to 9.4 days. Length of stay in public hospitals has fluctuated in a narrow range from 9 to 10.4 days but fell by more than a day between 1990 and 1994. Length of stay in proprietary hospitals shows the least change, about a day over the period, and remains relatively high (11.3 days). In both proprietary and public hospitals the changes in length of stay during the 6-year period are not statistically significant but there seems to be a narrowing in length of stay variation among types of hospitals over time.Figure 1.Average length of stay in general hospitals for psychiatric discharges by hospital ownership. Lengths of stay of less than 1 day are recoded to equal 1 day. Means (approximate SEs) are presented.Between 1988 and 1994, hospital admissions of 5 days or less significantly increased (z=3.22; P≤.01), from 40% to 45% of all discharges during the period studied. As shown in Figure 2, in each year public hospitals admitted a significantly larger proportion of their patients for such short stays than other types of hospitals, although in 1994 the difference between public hospitals and private nonprofit hospitals is not significant. Over time, the proportion of discharges following such short admissions modestly increased in public hospitals, from 51% to 53%. There was a larger, significant increase (z=5.82, P≤.001) in the proportion of short hospital stays in nonprofit private hospitals, from 37% in 1988 to 46% in 1994. In contrast, the proportion of such cases significantly declined (z=−2.76, P≤.01) in proprietary hospitals, from 43% in 1988 to 36% in 1994.Figure 2.Percentage of psychiatric discharges to the community after hospitalization of 5 days or less by hospital ownership. Analysis included only discharges to the community (routine or against medical advice). Percentages (approximate SEs) are presented. Superscripts denote significant differences (P≤.05) between groups: a, different from public; b, different from private nonprofit; and c, different from proprietary. Asterisk indicates relative SE of the numerator greater than 10% (<12.5%).Detailed data on length of stay for patients in mental hospitals are no longer reported by the Center for Mental Health Services; however, the American Hospital Association reports that the average length of stay in mental hospitals declined from 75 to 56 days between 1988 and 1992.TOTAL DAYS OF CAREThe number of inpatient days provided in general hospitals declined in 1990 and 1992 but in 1994 was moderately (7%) higher than in 1988, although there was not a significant linear trend during the 6 years (Figure 3). In each year private nonprofit hospitals provide more than 70% of the total days of care; however, the increase of approximately 1.7 million days in this 6-year period is not statistically significant. In contrast, proprietary hospitals experienced a significant 19% growth (z=2.15; P≤.01) in the number of days of care provided, while public general hospitals provided 39% fewer total days of care in 1994 than they did in 1988 (z=−3.55; P≤.001).Figure 3.Total psychiatric days of care (in millions) in general hospitals by hospital ownership. Total days (approximate SEs) are presented. Superscripts denote significant differences (P≤.05) between groups: a, different from public; b, different from private nonprofit; and c, different from proprietary.In comparison, between 1988 and 1994 the total days of care provided in mental hospitals declined by approximately 28% (data not shown). This occurred in all 3 types of hospitals, but the decline was largest (30%) in public mental hospitals. Overall, mental hospitals provided more than three quarters of all days of care each year but have also experienced a substantial reduction of 12.5 million inpatient days during this period. This reduction far exceeds the net increase of 1.2 million days in general hospitals during this 6-year period.PROPORTION OF PERSONS WITH SMIAs presented in Table 3, the rate of discharge for SMI has significantly increased in general hospitals from approximately 196 to 314 discharges per 100000 population. The largest increase was experienced in private nonprofit hospitals, where rates of discharges increased by almost 90%. While the rate of discharge for SMI in proprietary hospitals is higher in 1994 than in 1988, the overall trend is not significant. In contrast, there was a significant decline in rate of discharges with SMI in public hospitals.Table 3. Rate of Discharges per 100000 Adult Population and Proportion of Discharges and Days of Care for Patients With a Severe Mental Illness in General Hospitals by Hospital Ownership*See table graphicIn all 3 types of hospitals, diagnoses of SMI constituted an increasing proportion of all psychiatric discharges between 1988 and 1994; however, the trend is statistically significant only in private nonprofit hospitals. Other analyses (data not shown) indicate that the SMI diagnoses that increased the most were schizophrenia in private nonprofit hospitals and major depression in proprietary hospitals.In all 3 types of hospitals, the proportion of total days of care provided to patients with SMI significantly increased between 1988 and 1994. The increase is largest in private nonprofit hospitals, where approximately 12% more of the total bed-days were provided to patients with SMI in 1994 than in 1988. While in each year a larger proportion of total days of care are provided to patients with SMI in public hospitals than in either type of private hospital, the differences between the private and the public sector have diminished over time.SOURCES OF PAYMENTPublic financing for psychiatric inpatient care in general hospitals increased significantly during the 1988 to 1994 period, while private financing significantly declined. As presented in Table 4, in 1994 only 1 in 4 days of psychiatric care in general hospitals was paid through private insurance and Medicare and Medicaid paid for 60% of all inpatient days. During the period studied, the proportion of days paid through private insurance significantly declined by 22% in proprietary hospitals, 17% in private nonprofit hospitals, and 5% in public hospitals. The burden of care for the uninsured continues to fall disproportionally on public hospitals, and there has been a modest decline in the number of days of care provided for the uninsured in private nonprofit hospitals. However, private nonprofit hospitals provided the majority of days of care to the uninsured; more than 1 million days in 1994 as compared with 165000 days in public hospitals. The proportion of days of care financed through Medicaid has significantly increased in all types of general hospitals, but the growth is especially striking in the private sector. In 1994, Medicaid was the largest expected payer in public hospitals and Medicare the largest expected payer in both types of private hospitals.Table 4. Percentage of Total Psychiatric Days of Care in General Hospitals for Each Payer Source by Hospital Ownership*See table graphicDISCHARGE STATUSParticularly striking in Figure 4is the growth in the proportion of patients transferred from proprietary hospitals that is evident in 1992 and accelerates in 1994. In 1994, proprietary hospitals transferred almost 18% more of their patients than they did in 1988.Figure 4.Percentage of psychiatric discharges transferred by hospital ownership. Discharge status was missing for 3.3% of discharges in 1988, 4.9% in 1990, 4.7% in 1992, and 6.5% in 1994. Percentages (approximate SEs) are presented. Superscripts denote significant (P≤.05) differences between groups: a, different from public; b, different from private nonprofit; and c, different from proprietary. Asterisk indicates relative SE of the numerator greater than 10% (<20.0%).We compared rates of transfer between proprietary and private nonprofit hospitals in 1994 where subgroup sizes allow reliable analysis. Transfer rates are higher in proprietary hospitals regardless of source of payment or diagnostic category. The largest difference between hospitals is for Medicaid enrollees and patients who are uninsured; proprietary hospitals transferred approximately 1 in 4 of such patients in 1994, compared with less then 1 in 10 in private nonprofit hospitals (data not shown). Proprietary hospitals transferred approximately 14% more of their patients with non-SMI and 17% more of their patients with SMI than did private nonprofit hospitals. The highest rate of transfer within the SMI category was for patients with a diagnosis of schizophrenia; in 1994 proprietary hospitals transferred about half of their patients with schizophrenia, compared with 19% transferred from private nonprofit hospitals.COMMENTManaged care and an increased focus on community alternatives to inpatient care would lead us to predict some reduction in general hospital discharges and a substantial reduction in bed-days. In contrast, we find a substantial increase over time in discharges and in the proportion of patients with severe disorders. Patterns of general hospital care, however, cannot be understood without considering trends in the traditional mental hospital sector and in public mental health policy.States have continued to downsize their public mental hospitals, to return long-term patients to community residences, and to transfer care to general hospitals. Many patients who previously would have received care in public mental hospitals now receive general hospital care. Inpatient episodes are typically short and focused on managing crises and stabilizing symptoms. Thus, it seems that patients who are less seriously ill are increasingly less likely to be admitted to hospital and are being replaced by publicly financed patients with more serious diagnoses.The evidence for this characterization is circumstantial because these surveys are limited, but the observations support this picture. Rates of discharge from general hospitals have increased markedly, while the decline in length of stay has meant that total bed-days have only increased modestly. Moreover, the increase of 1.2 million bed-days in general hospitals between 1988 and 1994 is small relative to the reduction of 12.5 million bed-days in mental hospitals during this same period. Many long-term patients who once occupied mental hospital beds now live in small community institutions such as nursing homes, intermediate care facilities, board and care homes, and supervised residences, using only occasional hospital care. Some states may also be recreating mental institutions in the community.Over time, the proportion of patients with SMI in general hospitals has increased. Such patients are increasingly concentrated in private nonprofit general hospitals although public general hospitals continue to provide care disproportionately for Medicaid enrollees, the uninsured, and for patients with SMI.The large proportion of patients who are discharged after very short admissions raises questions about quality of care. Short hospital stays may be disadvantageous to the degree they provide inadequate preparation for discharge and increase readmission rates.However, there is some indication that very short hospital stays may be effective when successful links to outpatient treatment are established.Given the increased proportion of discharges following such short admissions, patient groups likely to benefit or be adversely affected need to be more carefully examined.Public facilities in the general hospital and mental hospital sectors are providing substantially less care than in prior years. Discharges for psychiatric disorders from public general hospitals declined by almost one third between 1988 and 1994. The largest decline, however, has been in discharges with less severe diagnoses, and over time public hospitals are providing a significantly greater proportion of all days of care to persons with SMI. These hospitals also continued to be disproportionately responsible for the most disadvantaged patients, when compared with hospitals in the private sectors.With the erosion of public-sector institutions, much concern has been expressed about the maintenance of the psychiatric safety net.Although in many cases the bed capacity in public general hospitals to care for persons with mental illness has not diminished, the increasing attractiveness of Medicaid enrollees to nonprofit general hospitals has left public hospitals with a disproportionate share of the most difficult and least renumerative patients, threatening financial insolvency. In private nonprofit hospitals the proportion of days of care financed through Medicaid almost doubled between 1988 and 1994. While proprietary general hospitals admit fewer Medicaid enrollees, the proportion of all bed-days financed through this program more than doubled in the 6 years studied. As more states put their Medicaid population into behavioral health programs, and as private insurance increasingly diverts hospital admissions and expedites discharge, these trends are likely to accelerate. Moreover, this trend provides some support for the hypothesis that increased competition reduces differences between the profit and nonprofit sectors,although we do not find a reduction over time in the different amounts of uncompensated care provided between these sectors.There are some promising signs in these developments. As private nonprofit facilities assume more of the burden, they help to establish a single class of hospital service for persons with psychiatric disorders that departs from the multiple tiers that characterize earlier patterns. Nonprofit hospitals now substantially serve Medicaid enrolles and provide considerable uncompensated care. Moreover, psychiatric inpatient care seems to be more targeted to persons with serious disorders. This burden is not distributed evenly among nonprofit institutions. If such institutions are to replace the public institutional safety net we will need to sharpen our capacity to identify and target such hospitals for public support.Inpatient care is the most expensive component of mental health services and accounts for most mental health expenditure. Ironically, we know very little about the content of care in these institutions, appropriate norms for inpatient stays, or the inpatient arrangements that results in the best outcomes and connection with needed aftercare services.Inpatient psychiatric care must be part of a balanced system of services with strong links to a broad array of services outside the hospital. Various studies suggest that linkages are commonly poor, and often patients with the most severe and persistent disorders do not receive the aftercare and community services they require, leading to exacerbations and homelessness.Although much attention has been focused on the threat of capitation and managed care for the treatment of mental illness, effective care management can provide opportunities to achieve better integration between inpatient interventions and community aftercare services.The proprietary sector is only a small component of inpatient care, but trends in patient transfers strongly suggest "dumping," consistent with the findings of another recent study.The high rate of transfer of Medicaid patients, the uninsured, and patients with schizophrenia requires closer scrutiny.The data analyses have important limitations. We do not include discharges from hospitals controlled by the Department of Veterans Affairs, although they play a minor role in inpatient psychiatric care (approximately 160000 discharges in 1994). More importantly, the data sources used are based on discharges, not individuals, and we cannot determine the number and types of patients with more than 1 discharge in any year. The 2 data sources used are not intended to be compatible and they include some duplication. We cannot differentiate between discharges from dedicated units and those from "scatter beds," although earlier studies have noted important differences in the populations treated and the services received.Also, we can not take account of changes in diagnostic practices, Medicare and Medicaid eligibility, and age structure over time, or use these data to examine patterns in delimited geographic areas. These limitations suggest the inadequacy of available data sources for tracking mental health service utilization and the need for improvement.LLBachrachThe state of the state mental hospital in 1996.Psychiatr Serv.1996;47:1071-1078.RWRedickMJWitkinJEAtayRWManderscheidHighlights of organized mental health services in 1992 and major national and state trends.In: Manderscheid RW, Sonnenschein MA, eds. Mental Health, United States, 1996. Washington, DC: US Dept of Health and Human Services, Public Health Service; 1996:90-137.JPBaeAssessing the need, use, and developments in mental health/substance abuse care.Health Care Financ Rev.1997;18:1-4.Health Care Financing AdministrationNational summary of Medicaid managed care programs and enrollment.Available from: http://www.hcfa.gov/medicaid/trends1.htm. Accessed August 1997.BSGillumEJGravesLJeanTrends in hospital utilization: United States, 1988-1992.Vital Health Stat 13.1996;124:1-71.EJGravesBGillumNational Hospital Discharge Survey: Annual Summary, 1994.Data from Vital and Health Statistics, No. 128.Not AvailableInternational Classification of Diseases, Ninth Revision, Clinical Modification.Washington, DC: Public Health Service, US Dept of Health and Human Services; 1988.US Bureau of the CensusUS population estimates by age, sex, race, and hispanic origin: 1980-1996.Available from: http://www.census.gov/population/estimates/nation. Accessed August 1997.JDNeedlemanDJCholletJLamphereHospital conversion trends.Health Aff (Millwood).1997;16:187-195.American Hospital AssociationAmerican Hospital Association Hospital Statistics, 1993-1994.Chicago, Ill: American Hospital Association; 1993.American Hospital AssociationAmerican Hospital Association Hospital Statistics, 1989-1990.Chicago, Ill: American Hospital Association; 1989.HRLambThe new state mental hospital in the community.Psychiatr Serv.1997;48:1307-1310.LApplebyPNDesaiDJLuchinRDGibbonsDRHedekerLength of stay and recidivism in schizophrenia: a study of public psychiatric hospital patients.Am J Psychiatry.1993;150:72-76.SESchneiderIMRossUltra-short hospitalizations for severely mentally ill patients.Psychiatr Serv.1996;47:137-138.RBaxterREMechanicThe status of local health care safety nets.Health Aff (Millwood).1997;16:7-23.MSchlesingerRDorwartCHooverSEpsteinCompetition and access to hospital services: evidence from psychiatric hospitals.Med Care.1997;35:974-992.DMechanicImproving Inpatient Psychiatric Treatment in an Era of Managed Care.San Francisco, Calif: Jossey-Bass; 1997. New Directions in Mental Health Services No. 73.CBoyerMeaningful linkage practices: challenges and opportunities.In: Mechanic D, ed. Improving Inpatient Psychiatric Treatment in an Era of Managed Care. San Francisco, Calif: Jossey-Bass; 1997:87-101. New Directions in Mental Health Services No. 73.DMechanicManaged Behavioral Health Care: Current Realities and Future Potential.San Francisco, Calif: Jossey-Bass; 1998. New Directions in Mental Health Services No. 78.MSchlesingerRDorwartCHooverSEpsteinThe determinants of dumping: a national study of economically motivated transfer involving mental health care.Health Serv Res.1997;32:561-590.CAKieslerCGSimpkinsThe Unnoticed Majority in Psychiatric Inpatient Care.New York, NY: Plenum Press; 1993.DMechanicDDavisPatterns of care in general hospitals for patients with psychiatric diagnoses: some findings and some cautions.Med Care.1990;28:1153-1164.Accepted for publication May 14, 1998.This article is part of a large collaborative effort funded by the Robert Wood Johnson Foundation, Princeton, NJ (Healthcare for Communities: The Alcohol, Drug and Mental Illness Tracking Study).We thank Kaila Simmons for her assistance with the graphics.Reprints: David Mechanic, PhD, Institute for Health, Health Care Policy and Aging Research, 30 College Ave, Rutgers University, New Brunswick, NJ 08901. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png JAMA Psychiatry American Medical Association

Changing Patterns of Psychiatric Inpatient Care in the United States, 1988-1994

Loading next page...
 
/lp/american-medical-association/changing-patterns-of-psychiatric-inpatient-care-in-the-united-states-qNGJyG0asY

References (24)

Publisher
American Medical Association
Copyright
Copyright 1998 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.
ISSN
2168-622X
eISSN
2168-6238
DOI
10.1001/archpsyc.55.9.785
Publisher site
See Article on Publisher Site

Abstract

Using data from the National Hospital Discharge Survey and the Inventory of Mental Health Organizations, this article examines national trends in psychiatric inpatient care from 1988 to 1994 in general hospitals and mental hospitals. We find that discharges with a primary diagnosis of mental illness in general hospitals increased from 1.4 to 1.9 million during this period. The total increase of 1.2 million days of care in general hospitals was small relative to the reduction of 12.5 million inpatient days in mental hospitals. General hospital discharges increased most in private nonprofit hospitals and declined substantially in public hospitals. Length of stay has fallen most substantially in private nonprofit hospitals. Public programs have increasingly replaced private insurance as the major source of payment. These observations suggest that psychiatric inpatient care in general hospitals can be characterized as a process in which patients who would have been clients of public mental hospitals in a prior period replace privately insured patients who, under managed care, are largely treated in community settings. Private nonprofit general hospitals increasingly treat publicly financed patients with more severe illnesses.In recent years the continuing reduction of resident populations in long-term mental hospitals, hospital closures and mergers, managed care, and an increasingly competitive marketplace have transformed the psychiatric inpatient sector. Resident populations in public psychiatric hospitals fell to less than 80000 in the 1990s.In contrast, the number of specialized psychiatric units in general hospitals increased from 664 to 1516 between 1970 and 1992 and the number of private mental hospitals more than tripled, with inpatient admissions quadrupling.Most Americans are now in behavioral health care programsand most persons with severe mental illness (SMI) reside in the community, commonly with Medicaid coverage. Medicaid managed care enrollment has been growing rapidly, increasing from less than 10% in 1991 to 48% in 1997.The consequences of these changes are poorly understood.In this article we examine how these changes in the health care system have shaped patterns of psychiatric inpatient care. We explore the hypothesis that managed care reduces inpatient care for less seriously ill persons in general hospitals and that these patients are replaced by persons with more severe disorders who in earlier periods received their care in public mental hospitals. By examining trends in relationship to ownership we also seek insights into how competition affects selection into different types of hospitals. We do this by examining changes in discharge numbers, rates and days of care, lengths of stay, proportions of inpatients with SMI, sources of payment, and discharge status.MATERIALS AND METHODSData for these analyses come from the National Hospital Discharge Survey (NHDS) and the Inventory of Mental Health Organizations and General Hospital Mental Health Services (hereafter referred to as "Inventory") for the period 1988 to 1994.The NHDS, described in detail elsewhere,is a multistage sample of discharges from nonfederal hospitals with an average length of stay of less than 30 days. Short-stay mental hospitals are eligible for inclusion in the sample, but the number included is small and represents approximately 13% of all psychiatric discharges (R. Pokras, written communication, August 1997). For all analyses, sample statistics are weighted to allow for national estimates.Analyses are limited to discharges with first-listed psychiatric diagnoses coded between 290 and 319 according to the International Classification of Diseases, 9th Revision, Clinical Modification.Severe mental illness includes a first-listed diagnosis of schizophrenia and related disorders (codes 295, 297-299), bipolar disorder (296.0, 296.1, 296.4-296.9), obsessive-compulsive disorder (300.3), or major depressive disorder with psychotic features (296.34, 296.24). Rates of discharge are calculated using US Census estimates of the adult, civilian, noninstitutionalized population on July 1 of each year.We distinguish between public hospitals, including those controlled by state and local governments; private nonprofit hospitals, including those operated by not-for-profit organizations and churches; and proprietary hospitals, including those operated for profit by individuals, corporations, or partnerships. Changes in hospital ownership are coded every 3 years; thus, the 1990 data do not accurately reflect ownership for those hospitals that changed control after 1988. The number of hospitals miscoded in 1990, however, is likely to be small because only approximately 1% of hospitals change ownership each year.Other variables include a continuous measure of length of stay (difference between admission and discharge date) and a categorical variable measuring short lengths of stay, defined here as 5 days or less. Total days of care is the sum of lengths of stay. Expected payer source is defined as the primary listed source of payment. Transfer rates are the ratio of the number of patients transferred to another institution at discharge to the total number of discharges for a particular group.The Inventory, a biennial national survey by the Center for Mental Health Services (Rockville, Md) of facilities that provide specialty mental health care, provides supplementary data and is described in detail elsewhere.The Inventory included very limited clinical information for each admission, and we use it only to estimate changes in the number and rates of discharges and total days of care from mental hospitals. Sampling overlap between the NHDS and the Inventory may in some cases increase aggregate estimates by 10% to 15% when data from these 2 samples are combined.Because the Inventory is based on surveys of the universe of hospitals, tests for statistical significance are not appropriate. We use the parameters provided by the National Center for Health Statistics to calculate approximate SEs for various estimates for the NHDS.We highlight when estimates are based on less than 60 cases or have a relative SE of more than 30%, because these estimates may be unreliable. Calculation of the SE of percentages assumes that the relative SE of the denominator is less than 5% of the estimate, or that the relative SEs of both the denominator and the numerator are less than 10%. Some subgroup analyses presented here violate this assumption. However, the SEs and tests of significance presented probably will be conservative, given that we cannot calculate exact SEs for the major variables. We recommend that attention be focused on the magnitude of substantive differences between groups, and not solely on whether the test statistic attains statistical significance.Statistical inferences about differences between types of hospitals are made through 2-tailed ttests. Analyses of changes over time used weighted least squares regression,testing whether there is a linear relationship between time (year of survey) and each of our measures.RESULTSNUMBER AND RATE OF DISCHARGEAs presented in Table 1, between 1988 and 1994 the number of discharges for a primary psychiatric diagnosis from general hospitals significantly increased by approximately 35%, from 1.4 to 1.9 million discharges. The rate of discharges also significantly increased during this period, from 785 to 996 discharges per 100000 adult population. With a modest decrease in discharges during the same period for other illnesses (1.4%), psychiatric discharges increased as a proportion of all discharges from 5.1% to 6.8%.Table 1. Psychiatric Discharges (in Millions) From General Hospitals and Rate per 100000 Adult Population by Hospital Ownership*See table graphicThe increase in the number and rate of discharges has been confined to the private sector. The largest growth in inpatient psychiatric care in general hospitals occurred in private nonprofit hospitals (53% increase), with a smaller (29%), though not statistically significant, increase in the number of discharges from proprietary hospitals. While significantly more discharges were from public hospitals than proprietary hospitals in 1988, in 1994 the opposite pattern emerges. Indeed, during these 6 years public hospitals experienced an almost one-third reduction in number of discharges. Rates of discharge per 100000 population follow this same pattern.During this same period, discharges from mental hospitals also increased, although the rate of discharge from these hospitals, unlike that from general hospitals, remained relatively stable (Table 2). Discharges from public mental hospitals declined by approximately 20%, while increasing by more than 27% in private nonprofit and proprietary hospitals. Combining discharges from general and mental hospitals shows an aggregate increase of more than half a million discharges in 1994 as compared with 1988. During this period, the proportion of all discharges accounted for by general hospitals increased from 68% to 73%.Table 2. Discharges From Mental Hospitals and Rate per 100000 Adult Population by Hospital Ownership*See table graphicLENGTH OF STAYIn the period 1988 to 1994, psychiatric length of stay in general hospitals declined significantly from 12.1 to 9.6 days. As shown in Figure 1, the steepest decline occurred in nonprofit general hospitals, where average length of stay fell from 12.6 to 9.4 days. Length of stay in public hospitals has fluctuated in a narrow range from 9 to 10.4 days but fell by more than a day between 1990 and 1994. Length of stay in proprietary hospitals shows the least change, about a day over the period, and remains relatively high (11.3 days). In both proprietary and public hospitals the changes in length of stay during the 6-year period are not statistically significant but there seems to be a narrowing in length of stay variation among types of hospitals over time.Figure 1.Average length of stay in general hospitals for psychiatric discharges by hospital ownership. Lengths of stay of less than 1 day are recoded to equal 1 day. Means (approximate SEs) are presented.Between 1988 and 1994, hospital admissions of 5 days or less significantly increased (z=3.22; P≤.01), from 40% to 45% of all discharges during the period studied. As shown in Figure 2, in each year public hospitals admitted a significantly larger proportion of their patients for such short stays than other types of hospitals, although in 1994 the difference between public hospitals and private nonprofit hospitals is not significant. Over time, the proportion of discharges following such short admissions modestly increased in public hospitals, from 51% to 53%. There was a larger, significant increase (z=5.82, P≤.001) in the proportion of short hospital stays in nonprofit private hospitals, from 37% in 1988 to 46% in 1994. In contrast, the proportion of such cases significantly declined (z=−2.76, P≤.01) in proprietary hospitals, from 43% in 1988 to 36% in 1994.Figure 2.Percentage of psychiatric discharges to the community after hospitalization of 5 days or less by hospital ownership. Analysis included only discharges to the community (routine or against medical advice). Percentages (approximate SEs) are presented. Superscripts denote significant differences (P≤.05) between groups: a, different from public; b, different from private nonprofit; and c, different from proprietary. Asterisk indicates relative SE of the numerator greater than 10% (<12.5%).Detailed data on length of stay for patients in mental hospitals are no longer reported by the Center for Mental Health Services; however, the American Hospital Association reports that the average length of stay in mental hospitals declined from 75 to 56 days between 1988 and 1992.TOTAL DAYS OF CAREThe number of inpatient days provided in general hospitals declined in 1990 and 1992 but in 1994 was moderately (7%) higher than in 1988, although there was not a significant linear trend during the 6 years (Figure 3). In each year private nonprofit hospitals provide more than 70% of the total days of care; however, the increase of approximately 1.7 million days in this 6-year period is not statistically significant. In contrast, proprietary hospitals experienced a significant 19% growth (z=2.15; P≤.01) in the number of days of care provided, while public general hospitals provided 39% fewer total days of care in 1994 than they did in 1988 (z=−3.55; P≤.001).Figure 3.Total psychiatric days of care (in millions) in general hospitals by hospital ownership. Total days (approximate SEs) are presented. Superscripts denote significant differences (P≤.05) between groups: a, different from public; b, different from private nonprofit; and c, different from proprietary.In comparison, between 1988 and 1994 the total days of care provided in mental hospitals declined by approximately 28% (data not shown). This occurred in all 3 types of hospitals, but the decline was largest (30%) in public mental hospitals. Overall, mental hospitals provided more than three quarters of all days of care each year but have also experienced a substantial reduction of 12.5 million inpatient days during this period. This reduction far exceeds the net increase of 1.2 million days in general hospitals during this 6-year period.PROPORTION OF PERSONS WITH SMIAs presented in Table 3, the rate of discharge for SMI has significantly increased in general hospitals from approximately 196 to 314 discharges per 100000 population. The largest increase was experienced in private nonprofit hospitals, where rates of discharges increased by almost 90%. While the rate of discharge for SMI in proprietary hospitals is higher in 1994 than in 1988, the overall trend is not significant. In contrast, there was a significant decline in rate of discharges with SMI in public hospitals.Table 3. Rate of Discharges per 100000 Adult Population and Proportion of Discharges and Days of Care for Patients With a Severe Mental Illness in General Hospitals by Hospital Ownership*See table graphicIn all 3 types of hospitals, diagnoses of SMI constituted an increasing proportion of all psychiatric discharges between 1988 and 1994; however, the trend is statistically significant only in private nonprofit hospitals. Other analyses (data not shown) indicate that the SMI diagnoses that increased the most were schizophrenia in private nonprofit hospitals and major depression in proprietary hospitals.In all 3 types of hospitals, the proportion of total days of care provided to patients with SMI significantly increased between 1988 and 1994. The increase is largest in private nonprofit hospitals, where approximately 12% more of the total bed-days were provided to patients with SMI in 1994 than in 1988. While in each year a larger proportion of total days of care are provided to patients with SMI in public hospitals than in either type of private hospital, the differences between the private and the public sector have diminished over time.SOURCES OF PAYMENTPublic financing for psychiatric inpatient care in general hospitals increased significantly during the 1988 to 1994 period, while private financing significantly declined. As presented in Table 4, in 1994 only 1 in 4 days of psychiatric care in general hospitals was paid through private insurance and Medicare and Medicaid paid for 60% of all inpatient days. During the period studied, the proportion of days paid through private insurance significantly declined by 22% in proprietary hospitals, 17% in private nonprofit hospitals, and 5% in public hospitals. The burden of care for the uninsured continues to fall disproportionally on public hospitals, and there has been a modest decline in the number of days of care provided for the uninsured in private nonprofit hospitals. However, private nonprofit hospitals provided the majority of days of care to the uninsured; more than 1 million days in 1994 as compared with 165000 days in public hospitals. The proportion of days of care financed through Medicaid has significantly increased in all types of general hospitals, but the growth is especially striking in the private sector. In 1994, Medicaid was the largest expected payer in public hospitals and Medicare the largest expected payer in both types of private hospitals.Table 4. Percentage of Total Psychiatric Days of Care in General Hospitals for Each Payer Source by Hospital Ownership*See table graphicDISCHARGE STATUSParticularly striking in Figure 4is the growth in the proportion of patients transferred from proprietary hospitals that is evident in 1992 and accelerates in 1994. In 1994, proprietary hospitals transferred almost 18% more of their patients than they did in 1988.Figure 4.Percentage of psychiatric discharges transferred by hospital ownership. Discharge status was missing for 3.3% of discharges in 1988, 4.9% in 1990, 4.7% in 1992, and 6.5% in 1994. Percentages (approximate SEs) are presented. Superscripts denote significant (P≤.05) differences between groups: a, different from public; b, different from private nonprofit; and c, different from proprietary. Asterisk indicates relative SE of the numerator greater than 10% (<20.0%).We compared rates of transfer between proprietary and private nonprofit hospitals in 1994 where subgroup sizes allow reliable analysis. Transfer rates are higher in proprietary hospitals regardless of source of payment or diagnostic category. The largest difference between hospitals is for Medicaid enrollees and patients who are uninsured; proprietary hospitals transferred approximately 1 in 4 of such patients in 1994, compared with less then 1 in 10 in private nonprofit hospitals (data not shown). Proprietary hospitals transferred approximately 14% more of their patients with non-SMI and 17% more of their patients with SMI than did private nonprofit hospitals. The highest rate of transfer within the SMI category was for patients with a diagnosis of schizophrenia; in 1994 proprietary hospitals transferred about half of their patients with schizophrenia, compared with 19% transferred from private nonprofit hospitals.COMMENTManaged care and an increased focus on community alternatives to inpatient care would lead us to predict some reduction in general hospital discharges and a substantial reduction in bed-days. In contrast, we find a substantial increase over time in discharges and in the proportion of patients with severe disorders. Patterns of general hospital care, however, cannot be understood without considering trends in the traditional mental hospital sector and in public mental health policy.States have continued to downsize their public mental hospitals, to return long-term patients to community residences, and to transfer care to general hospitals. Many patients who previously would have received care in public mental hospitals now receive general hospital care. Inpatient episodes are typically short and focused on managing crises and stabilizing symptoms. Thus, it seems that patients who are less seriously ill are increasingly less likely to be admitted to hospital and are being replaced by publicly financed patients with more serious diagnoses.The evidence for this characterization is circumstantial because these surveys are limited, but the observations support this picture. Rates of discharge from general hospitals have increased markedly, while the decline in length of stay has meant that total bed-days have only increased modestly. Moreover, the increase of 1.2 million bed-days in general hospitals between 1988 and 1994 is small relative to the reduction of 12.5 million bed-days in mental hospitals during this same period. Many long-term patients who once occupied mental hospital beds now live in small community institutions such as nursing homes, intermediate care facilities, board and care homes, and supervised residences, using only occasional hospital care. Some states may also be recreating mental institutions in the community.Over time, the proportion of patients with SMI in general hospitals has increased. Such patients are increasingly concentrated in private nonprofit general hospitals although public general hospitals continue to provide care disproportionately for Medicaid enrollees, the uninsured, and for patients with SMI.The large proportion of patients who are discharged after very short admissions raises questions about quality of care. Short hospital stays may be disadvantageous to the degree they provide inadequate preparation for discharge and increase readmission rates.However, there is some indication that very short hospital stays may be effective when successful links to outpatient treatment are established.Given the increased proportion of discharges following such short admissions, patient groups likely to benefit or be adversely affected need to be more carefully examined.Public facilities in the general hospital and mental hospital sectors are providing substantially less care than in prior years. Discharges for psychiatric disorders from public general hospitals declined by almost one third between 1988 and 1994. The largest decline, however, has been in discharges with less severe diagnoses, and over time public hospitals are providing a significantly greater proportion of all days of care to persons with SMI. These hospitals also continued to be disproportionately responsible for the most disadvantaged patients, when compared with hospitals in the private sectors.With the erosion of public-sector institutions, much concern has been expressed about the maintenance of the psychiatric safety net.Although in many cases the bed capacity in public general hospitals to care for persons with mental illness has not diminished, the increasing attractiveness of Medicaid enrollees to nonprofit general hospitals has left public hospitals with a disproportionate share of the most difficult and least renumerative patients, threatening financial insolvency. In private nonprofit hospitals the proportion of days of care financed through Medicaid almost doubled between 1988 and 1994. While proprietary general hospitals admit fewer Medicaid enrollees, the proportion of all bed-days financed through this program more than doubled in the 6 years studied. As more states put their Medicaid population into behavioral health programs, and as private insurance increasingly diverts hospital admissions and expedites discharge, these trends are likely to accelerate. Moreover, this trend provides some support for the hypothesis that increased competition reduces differences between the profit and nonprofit sectors,although we do not find a reduction over time in the different amounts of uncompensated care provided between these sectors.There are some promising signs in these developments. As private nonprofit facilities assume more of the burden, they help to establish a single class of hospital service for persons with psychiatric disorders that departs from the multiple tiers that characterize earlier patterns. Nonprofit hospitals now substantially serve Medicaid enrolles and provide considerable uncompensated care. Moreover, psychiatric inpatient care seems to be more targeted to persons with serious disorders. This burden is not distributed evenly among nonprofit institutions. If such institutions are to replace the public institutional safety net we will need to sharpen our capacity to identify and target such hospitals for public support.Inpatient care is the most expensive component of mental health services and accounts for most mental health expenditure. Ironically, we know very little about the content of care in these institutions, appropriate norms for inpatient stays, or the inpatient arrangements that results in the best outcomes and connection with needed aftercare services.Inpatient psychiatric care must be part of a balanced system of services with strong links to a broad array of services outside the hospital. Various studies suggest that linkages are commonly poor, and often patients with the most severe and persistent disorders do not receive the aftercare and community services they require, leading to exacerbations and homelessness.Although much attention has been focused on the threat of capitation and managed care for the treatment of mental illness, effective care management can provide opportunities to achieve better integration between inpatient interventions and community aftercare services.The proprietary sector is only a small component of inpatient care, but trends in patient transfers strongly suggest "dumping," consistent with the findings of another recent study.The high rate of transfer of Medicaid patients, the uninsured, and patients with schizophrenia requires closer scrutiny.The data analyses have important limitations. We do not include discharges from hospitals controlled by the Department of Veterans Affairs, although they play a minor role in inpatient psychiatric care (approximately 160000 discharges in 1994). More importantly, the data sources used are based on discharges, not individuals, and we cannot determine the number and types of patients with more than 1 discharge in any year. The 2 data sources used are not intended to be compatible and they include some duplication. We cannot differentiate between discharges from dedicated units and those from "scatter beds," although earlier studies have noted important differences in the populations treated and the services received.Also, we can not take account of changes in diagnostic practices, Medicare and Medicaid eligibility, and age structure over time, or use these data to examine patterns in delimited geographic areas. These limitations suggest the inadequacy of available data sources for tracking mental health service utilization and the need for improvement.LLBachrachThe state of the state mental hospital in 1996.Psychiatr Serv.1996;47:1071-1078.RWRedickMJWitkinJEAtayRWManderscheidHighlights of organized mental health services in 1992 and major national and state trends.In: Manderscheid RW, Sonnenschein MA, eds. Mental Health, United States, 1996. Washington, DC: US Dept of Health and Human Services, Public Health Service; 1996:90-137.JPBaeAssessing the need, use, and developments in mental health/substance abuse care.Health Care Financ Rev.1997;18:1-4.Health Care Financing AdministrationNational summary of Medicaid managed care programs and enrollment.Available from: http://www.hcfa.gov/medicaid/trends1.htm. Accessed August 1997.BSGillumEJGravesLJeanTrends in hospital utilization: United States, 1988-1992.Vital Health Stat 13.1996;124:1-71.EJGravesBGillumNational Hospital Discharge Survey: Annual Summary, 1994.Data from Vital and Health Statistics, No. 128.Not AvailableInternational Classification of Diseases, Ninth Revision, Clinical Modification.Washington, DC: Public Health Service, US Dept of Health and Human Services; 1988.US Bureau of the CensusUS population estimates by age, sex, race, and hispanic origin: 1980-1996.Available from: http://www.census.gov/population/estimates/nation. Accessed August 1997.JDNeedlemanDJCholletJLamphereHospital conversion trends.Health Aff (Millwood).1997;16:187-195.American Hospital AssociationAmerican Hospital Association Hospital Statistics, 1993-1994.Chicago, Ill: American Hospital Association; 1993.American Hospital AssociationAmerican Hospital Association Hospital Statistics, 1989-1990.Chicago, Ill: American Hospital Association; 1989.HRLambThe new state mental hospital in the community.Psychiatr Serv.1997;48:1307-1310.LApplebyPNDesaiDJLuchinRDGibbonsDRHedekerLength of stay and recidivism in schizophrenia: a study of public psychiatric hospital patients.Am J Psychiatry.1993;150:72-76.SESchneiderIMRossUltra-short hospitalizations for severely mentally ill patients.Psychiatr Serv.1996;47:137-138.RBaxterREMechanicThe status of local health care safety nets.Health Aff (Millwood).1997;16:7-23.MSchlesingerRDorwartCHooverSEpsteinCompetition and access to hospital services: evidence from psychiatric hospitals.Med Care.1997;35:974-992.DMechanicImproving Inpatient Psychiatric Treatment in an Era of Managed Care.San Francisco, Calif: Jossey-Bass; 1997. New Directions in Mental Health Services No. 73.CBoyerMeaningful linkage practices: challenges and opportunities.In: Mechanic D, ed. Improving Inpatient Psychiatric Treatment in an Era of Managed Care. San Francisco, Calif: Jossey-Bass; 1997:87-101. New Directions in Mental Health Services No. 73.DMechanicManaged Behavioral Health Care: Current Realities and Future Potential.San Francisco, Calif: Jossey-Bass; 1998. New Directions in Mental Health Services No. 78.MSchlesingerRDorwartCHooverSEpsteinThe determinants of dumping: a national study of economically motivated transfer involving mental health care.Health Serv Res.1997;32:561-590.CAKieslerCGSimpkinsThe Unnoticed Majority in Psychiatric Inpatient Care.New York, NY: Plenum Press; 1993.DMechanicDDavisPatterns of care in general hospitals for patients with psychiatric diagnoses: some findings and some cautions.Med Care.1990;28:1153-1164.Accepted for publication May 14, 1998.This article is part of a large collaborative effort funded by the Robert Wood Johnson Foundation, Princeton, NJ (Healthcare for Communities: The Alcohol, Drug and Mental Illness Tracking Study).We thank Kaila Simmons for her assistance with the graphics.Reprints: David Mechanic, PhD, Institute for Health, Health Care Policy and Aging Research, 30 College Ave, Rutgers University, New Brunswick, NJ 08901.

Journal

JAMA PsychiatryAmerican Medical Association

Published: Sep 1, 1998

There are no references for this article.